A REVISION OF THE ATOMIC WEIGHT OF ARSENIC. 



8l 



The filtrate and wash- waters were evaporated to small bulk. The precipitat- 

 ing flask was rinsed with ammonia, and the rinsing was added to the evapo- 

 rated filtrate and wash- water. Then the . solution was diluted to definite 

 volume, and the silver content was determined by comparison with standard 

 silver solutions in the nephelometer. 



The operations of precipitating and collecting the silver halides were all car- 

 ried out in a room lighted only with ruby light. 



INSOLUBLE RESIDUE. 



All the specimens of silver arsenate, after being heated at 250° C, when dis- 

 solved in dilute nitric acid, were found to contain a small amount of insoluble 

 residue, which would dissolve only in rather concentrated nitric acid. Although 

 the proportion of this residue was apparently increased by exposure to light, 

 specimens of the arsenate which had been prepared wholly in the dark room 

 were not free from it. No process of purification to which the soluble arsenate 

 used in the preparation of the silver arsenate was subjected seemed to have the 

 slightest effect upon the proportion of insoluble matter. A similar phenomenon 

 was met by Dr. Grinnell Jones in the preparation of silver phosphate (page 178). 



Although the amount of this residue in one gram of silver arsenate which 

 had been treated as in the analyses for silver was not over 0.00005 g^^i., it was 

 important to determine its silver content. This was done in three cases in 

 which the proportion of residue had been purposely increased as much as pos- 

 sible by exposure to light. The arsenate was dissolved in dilute nitric acid, 

 and the residue was collected upon a weighed platinum Gooch crucible, the 

 detached asbestos shreds being carefully determined by filtration upon a filter 

 paper. The weight of residue was found by reweighing the crucible. After 

 the residue had been dissolved in concentrated nitric acid and the solution had 

 been diluted to definite volume, the silver content of the solution was ascer- 

 tained by comparison with standard silver solutions in a nephelometer. 



The first of the above determinations was made with a sample of silver arse- 

 nate which had been exposed to bright light inside a desiccator for a month. 



